Data Visualization

Matplotlib

Matplotlib is an important library for creating static, animated, and interactive figure visualizations in Python.

• figsize: Defines the width and height of the figure in inches. The first value in paranthesis is width,and the second one is height.
• plt.xlabel() : Sets the label for the x-axis.
• plt.ylabel() : Sets the label for the y-axis.
• plt.title() : Specifies title of the visualization.
• plt.show() : Displays figures. If we do not write this code, the graph will be generated, however, would not be displayed on screen.

Seaborn

Seaborn is a library that uses Matplotlib underneath to plot graphs. It provides a high-level interface for drawing attractive and informative statistical graphics.

# Import Matplotlib ve Seaborn libraries

import matplotlib.pyplot as plt
import seaborn as sns


# Import pandas portion of the code to tell Python to bring the pandas data analysis library into our current environment.

import pandas as pd

df = pd.read_csv('credit_card_fixed.csv')

# Check the number of rows, number of columns of data set

df.shape

(451, 8)

#Access the first n (which is 10) rows of a data frame (df) of series

df.head(n = 10)

	have_credit_card	card_no	user_id	user_gender	user_age	user_city	total_amount	card_delivery_date
0	Yes	9871326599442434	1	Female	51	Edirne	3466	2022-01-03
1	Yes	5866643196858919	2	Male	33	İstanbul	2770	2022-01-22
2	Yes	5934969821915385	4	Female	45	İzmir	3990	2022-01-29
3	Yes	4583682992593297	5	Female	33	İstanbul	1876	2022-02-18
4	Yes	9834784352689854	6	Male	41	Ankara	4797	2022-02-07
5	Yes	1589797529681697	7	Male	55	Kocaeli	5164	2022-02-27
6	Yes	8976879654146946	8	Female	43	Amasya	4296	2022-01-13
7	Yes	4149817943357734	9	Male	26	Ankara	2099	2022-02-15
8	Yes	4448553446993154	10	Female	43	İzmir	5798	2022-01-18
9	Yes	9339661937631297	11	Male	51	Antalya	6595	2022-01-19

# Check the statistical descriptions out of the data 

df.describe()

	card_no	user_id	user_age	total_amount
count	4.510000e+02	451.000000	451.000000	451.000000
mean	5.574467e+15	249.853659	37.935698	3548.472284
std	2.631093e+15	144.039611	11.903233	1651.923984
min	1.112777e+15	1.000000	18.000000	725.000000
25%	3.377820e+15	122.500000	28.000000	2154.500000
50%	5.497366e+15	253.000000	37.000000	3295.000000
75%	7.849892e+15	374.500000	48.500000	4661.500000
max	9.982414e+15	500.000000	59.000000	8555.000000

Grouping Data in Python

groupby function is used to split the data into groups based on some criteria such as mean, min, max value, standard deviation. It is stated by agg which is an alias for aggregate.

# Age averages of users based on cities they live

df.groupby('user_city').agg({'user_age':'mean'})

	user_age
user_city
Amasya	37.605263
Ankara	37.416667
Antalya	40.146341
Bursa	40.523810
Edirne	38.742857
Kocaeli	37.382353
İstanbul	36.868852
İzmir	37.558140

# Max and min ages based on city names 

df.groupby('user_city').agg({'user_age': 'min'})

	user_age
user_city
Amasya	18
Ankara	18
Antalya	18
Bursa	20
Edirne	18
Kocaeli	18
İstanbul	18
İzmir	20

EXAMPLE - 1

Find the average age of the credit card users based on their genders.

df.groupby('user_gender').agg({'user_age': 'mean'})

	user_age
user_gender
Female	38.252174
Male	37.606335

Data Visualization Techniques

• Data visualization is the process of creating graphical representations of information.
• It allows us to gain insight into their vast amounts of data.

There are a number of ways to visualize data.

Here are some important data visualization techniques to know:

• Bar Chart, Pie Chart, Scatter Plot, Line Chart, Histogram, Heat Map, Box and Whisker Plot, Waterfall Chart, Area Chart etc.

1. Bar Chart

• It presents categorical data with rectangular bars with lengths proportional to the values that they represent.

• One axis of the plot shows the specific categories being compared, and the other axis represents a measured value.

Average credit card expenses based on cities

• Find the average credit card expenses based on cities.

df.groupby('user_city').agg({'total_amount':'mean'})

	total_amount
user_city
Amasya	3236.657895
Ankara	3815.385417
Antalya	3715.487805
Bursa	3426.214286
Edirne	3576.371429
Kocaeli	3302.647059
İstanbul	3453.893443
İzmir	3628.302326

df.groupby('user_city').agg({'total_amount':'mean'}).reset_index()

	user_city	total_amount
0	Amasya	3236.657895
1	Ankara	3815.385417
2	Antalya	3715.487805
3	Bursa	3426.214286
4	Edirne	3576.371429
5	Kocaeli	3302.647059
6	İstanbul	3453.893443
7	İzmir	3628.302326

#matplotlib

plt.figure(figsize = (12,6))

plt.bar(data = df.groupby('user_city').agg({'total_amount':'mean'}).reset_index(), 
        x = 'user_city',
        height = 'total_amount')
    
plt.show()

# Creating bar chart in seaborn

plt.figure(figsize = (12,6))
sns.barplot(data = df.groupby('user_city').agg({'total_amount':'mean'}).reset_index(), 
        y ='user_city',
        x ='total_amount')
plt.show()

# Find total money they spend based on cities
#You can write max or min instead of sum, and see them on bar chart

plt.figure(figsize = (12,6))
sns.barplot(data = df.groupby('user_city').agg({'total_amount':'sum'}).reset_index(), 
        y = 'user_city',
        x = 'total_amount')
plt.show()

Seaborn Styles

There are multiple options for the background in seaborn such as;

- sns.set_style('darkgrid')

- sns.set_style('whitegrid')

- sns.set_style('dark')

- sns.set_style('white')

sns.set_style('darkgrid')
#sns.set_style('whitegrid')
#sns.set_style('dark')
# sns.set_style('white')

## seaborn

plt.figure(figsize = (12,6))
sns.barplot(data = df.groupby('user_city').agg({'total_amount':'sum'}).reset_index(), 
        y = 'user_city',
        x = 'total_amount')
plt.show()

EXAMPLE - 2

Create a bar chart that shows credit card total amount they spent vs user gender with mathplot and seaborn libraries.

df.head()

	have_credit_card	card_no	user_id	user_gender	user_age	user_city	total_amount	card_delivery_date
0	Yes	9871326599442434	1	Female	51	Edirne	3466	2022-01-03
1	Yes	5866643196858919	2	Male	33	İstanbul	2770	2022-01-22
2	Yes	5934969821915385	4	Female	45	İzmir	3990	2022-01-29
3	Yes	4583682992593297	5	Female	33	İstanbul	1876	2022-02-18
4	Yes	9834784352689854	6	Male	41	Ankara	4797	2022-02-07

plt.figure(figsize = (12,6))
sns.barplot(data = df.groupby('user_gender').agg({'total_amount':'mean'}).reset_index(), 
        y='user_gender',
        x='total_amount')
plt.show()

2. Pie Chart

• Pie chart is generally used when trying to compare the catagorical data.
• Pie charts are ideal for giving a quick idea of the proportional distribution of data to the reader.
• plt.pie is used for creating pie chart,
• It is of two main parameters as x and labels.

Number of Observations of Cities

• Plot the number of credit card users based on cities.

df.groupby('user_city').agg({'user_id': 'count'}).reset_index()

	user_city	user_id
0	Amasya	38
1	Ankara	96
2	Antalya	41
3	Bursa	42
4	Edirne	35
5	Kocaeli	34
6	İstanbul	122
7	İzmir	43

plt.figure(figsize = (14,6))

plt.pie(data = df.groupby('user_city').agg({'user_id': 'count'}).reset_index(),
       x = 'user_id',
       labels = 'user_city')

plt.title('Credit card users based on cities;', size = 12, color = 'green')
plt.show()

Seaborn "Countplot" has the same function as "Pie Chart"!

# Color Palettes

sns.countplot(data = df,
             y = 'user_city',
            # color ='pink'
             palette = 'Set1')

plt.show()

EXAMPLE - 3

• Show the user numbers with Countplot or Pie Chart based on genders.

# Color Palettes

sns.countplot(data = df,
             x ='user_gender',
             color = 'green')
plt.title('Number of users based on genders', size = 12)
plt.show()

3. Scatter Plot

• If you have two variables that pair well together, plotting them on a scatter diagram is a great way to view their relationship and see if it's a positive or negative correlation.
• Even if we cannot understand the reason of the relationship, we can understand if there is a relationship or how strong the it is.
• plt.scatter is used to create a scatter plot with Matplotlib.

plt.scatter(data = df,
           x = 'user_age',
           y = 'total_amount')


plt.title('The relationship between the user age and credit card expenses', size = 14)
plt.xlabel('User age',size = 11)
plt.ylabel('Credit card expenses',size = 11)

plt.show()

4. Line Chart

• Line charts are used to represent quantitative data collected over a specific subject and a specific time interval.
• Generally we write dates/days on x-axis and quantitative data on y-axis
• plt.plot is used to create a scatter plot with Matplotlib.

Show the number of users who received credit cards on the basis of dates with a line chart

df['card_delivery_date'].value_counts()

2022-01-23    14
2022-01-25    13
2022-02-18    13
2022-01-06    13
2022-01-21    12
2022-01-01    12
2022-01-30    12
2022-01-11    11
2022-01-12    10
2022-01-03    10
2022-02-03    10
2022-01-28    10
2022-02-19    10
2022-02-08    10
2022-01-10    10
2022-01-17     9
2022-01-16     9
2022-02-02     9
2022-01-31     8
2022-02-25     8
2022-02-11     8
2022-02-27     8
2022-02-16     8
2022-01-04     8
2022-02-17     8
2022-01-18     8
2022-01-08     8
2022-01-20     7
2022-02-23     7
2022-02-04     7
2022-01-14     7
2022-01-05     7
2022-01-29     7
2022-01-22     7
2022-02-14     7
2022-02-07     7
2022-01-27     7
2022-01-09     7
2022-02-10     6
2022-01-15     6
2022-02-22     6
2022-01-19     6
2022-02-28     6
2022-02-12     6
2022-02-21     6
2022-02-06     6
2022-02-13     6
2022-02-15     6
2022-02-01     5
2022-02-05     5
2022-02-09     5
2022-01-07     5
2022-01-24     4
2022-01-02     4
2022-02-24     4
2022-01-26     4
2022-02-20     4
2022-02-26     3
2022-01-13     2
Name: card_delivery_date, dtype: int64

df['card_delivery_date'].dtype

dtype('O')

#Convert it to date format
#type(pd.to_datetime('2022-01-03'))

pd.to_datetime('2022-01-03')

Timestamp('2022-01-03 00:00:00')

df['card_delivery_date'] = pd.to_datetime(df['card_delivery_date'])
df

	have_credit_card	card_no	user_id	user_gender	user_age	user_city	total_amount	card_delivery_date
0	Yes	9871326599442434	1	Female	51	Edirne	3466	2022-01-03
1	Yes	5866643196858919	2	Male	33	İstanbul	2770	2022-01-22
2	Yes	5934969821915385	4	Female	45	İzmir	3990	2022-01-29
3	Yes	4583682992593297	5	Female	33	İstanbul	1876	2022-02-18
4	Yes	9834784352689854	6	Male	41	Ankara	4797	2022-02-07
...	...	...	...	...	...	...	...	...
446	Yes	2498987169455237	495	Female	43	Antalya	5895	2022-01-11
447	Yes	8483732676742838	496	Female	50	İzmir	5930	2022-01-18
448	Yes	5169329597692651	497	Male	52	Kocaeli	2675	2022-02-06
449	Yes	2676259853575781	499	Male	45	Antalya	2687	2022-02-04
450	Yes	7564466124434127	500	Male	38	Ankara	5233	2022-01-28

451 rows × 8 columns

card_df = df.groupby('card_delivery_date').agg({'user_id':'count'}).reset_index()

card_df

	card_delivery_date	user_id
0	2022-01-01	12
1	2022-01-02	4
2	2022-01-03	10
3	2022-01-04	8
4	2022-01-05	7
5	2022-01-06	13
6	2022-01-07	5
7	2022-01-08	8
8	2022-01-09	7
9	2022-01-10	10
10	2022-01-11	11
11	2022-01-12	10
12	2022-01-13	2
13	2022-01-14	7
14	2022-01-15	6
15	2022-01-16	9
16	2022-01-17	9
17	2022-01-18	8
18	2022-01-19	6
19	2022-01-20	7
20	2022-01-21	12
21	2022-01-22	7
22	2022-01-23	14
23	2022-01-24	4
24	2022-01-25	13
25	2022-01-26	4
26	2022-01-27	7
27	2022-01-28	10
28	2022-01-29	7
29	2022-01-30	12
30	2022-01-31	8
31	2022-02-01	5
32	2022-02-02	9
33	2022-02-03	10
34	2022-02-04	7
35	2022-02-05	5
36	2022-02-06	6
37	2022-02-07	7
38	2022-02-08	10
39	2022-02-09	5
40	2022-02-10	6
41	2022-02-11	8
42	2022-02-12	6
43	2022-02-13	6
44	2022-02-14	7
45	2022-02-15	6
46	2022-02-16	8
47	2022-02-17	8
48	2022-02-18	13
49	2022-02-19	10
50	2022-02-20	4
51	2022-02-21	6
52	2022-02-22	6
53	2022-02-23	7
54	2022-02-24	4
55	2022-02-25	8
56	2022-02-26	3
57	2022-02-27	8
58	2022-02-28	6

plt.figure(figsize = (12,6))

sns.lineplot(data = card_df,
            x = 'card_delivery_date',
            y = 'user_id')

plt.title('The number of user who recevied the credit cards based on dates', size = 12)

plt.show()

5. Histogram

• Histogram is used when the data are numerical.
• They display the shape of the data's frequency distribution, especially when determining whether the output of a process is distributed approximately normally.

Show the distribution of total expenditure with histogram

plt.figure(figsize = (12,6))
sns.histplot(data = df,
            x = 'total_amount')

plt.show()

EXAMPLE - 4

Show age distribution with histogram.

plt.figure(figsize = (8,5))
sns.histplot(data = df, x = 'user_age')

plt.show()

Plotly

• Plotly is a sophisticated data visualization library that is suited for creating elaborate plots efficiently.
• It is an open source tool, and has 40 different visualization package.
• It has 2 main parameters as data and layout.
• Data is the part where we specify the visualization type and data.
• Layout is the part where we specify the graph features such as titles, color, size, theme etc.

!pip install plotly

import plotly.graph_objs as go

from plotly.offline import iplot

# Total spending amaounts based on the cities where users live 

bar_df = df.groupby('user_city').agg({'total_amount' : 'sum'}).reset_index()

bar = go.Bar(x = bar_df['user_city'],
      y = bar_df['total_amount'])

fig = go.Figure(data = bar)

iplot(fig)

scatter = go.Scatter(x = df['user_age'],
          y = df['total_amount'],
                   mode ='markers')

fig = go.Figure(data = scatter)
iplot(fig)

scatter = go.Scatter(x = df['user_age'],
          y = df['total_amount'],
                   mode = 'markers',
                   marker = dict(size = 8,
                               color = 'blue',
                               opacity = 0.65)
                    )

layout = go.Layout(title = dict(text = 'The relationship between user age and total expenditure'),
                   xaxis = dict(title = 'Age'),
                   yaxis = dict(title = 'How much they spent'))

fig = go.Figure(data = scatter, layout = layout)

iplot(fig)

Relationship of Age-Total Expenditure Based on Genders

male = df.loc[df['user_gender'] == 'Male']
female = df.loc[df['user_gender'] == 'Female']


male_scatter = go.Scatter(x = male['user_age'],
                        y = male['total_amount'],
                        mode = 'markers',
                        name = 'Male')

female_scatter = go.Scatter(x = female['user_age'],
                        y = female['total_amount'],
                        mode ='markers',
                        name ='Feale')

gender_layout = go.Layout(title = dict(text = 'Age-Total Expenditure Relationship Based on Genders'),
                         xaxis = dict(title ='Age'),
                         yaxis = dict(title = 'Total Expenditure'),
                         template = 'plotly_dark')                          
                         
fig = go.Figure (data = [male_scatter, female_scatter], layout = gender_layout)
                         
iplot(fig)
                         

FINAL PROJECT

from plotly.subplots import make_subplots

def city_analysis():
    
    city = str(input('Enter a city: '))
    
    if city not in df['user_city'].unique():
        
        print('The city you enter is not on our list.. Sorry :(')
        
    else:
    
        city_data = df.loc[df['user_city'] == city]
        
        bar_df = city_data.groupby('user_gender').agg({'total_amount': 'mean'}).reset_index()
        
        fig = make_subplots(rows = 1,
                            cols = 2,
                            subplot_titles = ['Average consumption based on gender',
                                              'Age & Consumption Relationship'])
        
        fig.add_trace(go.Bar(x = bar_df['user_gender'],
                             y = bar_df['total_amount'],
                             showlegend = False),
                      
                      row = 1, col = 1)
        
        fig.add_trace(go.Scatter(x = city_data['total_amount'],
                                y = city_data['user_age'],
                                showlegend = False,
                                mode = 'markers'),
                     
                     row = 1, col = 2)
        
        fig.update_layout(title = dict(text = city + ' Credit Card Consumption Analysis ',
                                      x = 0.5),
                         xaxis = dict(title = 'Gender'),
                         yaxis = dict(title = 'Total Expenditure'),
                         xaxis2 = dict(title = 'Total Expenditure'),
                         yaxis2 = dict(title = 'Age'),
                         template = 'plotly_dark')
        
        iplot(fig)

city_analysis()

Enter a city: İstanbul

city_analysis()

Enter a city: Antalya

city_analysis()

Enter a city: Burdur
The city you enter is not on our list.. Sorry :(